Were We Wrong? 2022 Predictions Revisited

Episode Summary

Executive Summary: Wes and Scott review their 2022 web development predictions, finding many were accurate: SvelteKit/Svelte grew, Rust, serverless/edge, Tailwind, Deno, TypeScript, and server-first development all gained traction. They also assess slower-moving areas like GraphQL, Temporal, and Rome, while discussing developer tools, payment platforms, CSS advancements, and a pair of practical "sick picks" (a label maker and Chipolo trackers).

Main Topics: 2022 prediction recap: overall accuracy: The hosts revisit last year’s predictions and broadly conclude they were highly accurate, especially around framework adoption, runtime trends, and tooling shifts. Svelte, SvelteKit, and framework ergonomics: They praise SvelteKit’s straightforward API, note Svelte’s version 1 release, and describe it as a mature, elegant full-stack framework with strong docs and a low-friction developer experience. JavaScript ecosystem trends: TypeScript, Rust, Deno, serverless: They discuss the growing dominance of TypeScript, broader Rust adoption, Deno’s maturation, and the continued rise of serverless and edge functions across frameworks. Developer tooling and next-gen workflows: They cover Copilot, Warp, Fig, GitHub Code Search, and improved editor/terminal experiences, arguing that devtools are finally becoming smarter and more integrated. CSS and UI primitives: They evaluate container queries, :has, aspect-ratio, Open UI, and Open Props, concluding that CSS is moving forward but browser support and timing still shape real-world adoption. Data fetching and backend APIs: GraphQL, tRPC, server comeback: They contrast GraphQL’s relative stagnation and complexity with the rise of simpler typed APIs like tRPC and a broader shift back toward server-side rendering and server-first architectures. Practical product tools and subscriptions: They discuss payment processors and creator platforms (Stripe, Braintree, Gumroad, Lemon Squeezy, ConvertKit), emphasizing cost, merchant fees, and the build-vs-buy tradeoff.

Key Arguments: SvelteKit succeeded because it offers a clean, complete, and intuitive full-stack API without unnecessary complexity. Next.js improved with nested routing, but the ecosystem is increasingly favoring framework-agnostic libraries and adapters instead of tightly coupled data layers. Rust’s type discipline and low-level performance benefits are influencing more of the JavaScript ecosystem, including Bun and other runtime tooling. Serverless and edge functions have become mainstream integration points for modern frameworks, not niche infrastructure. Tailwind’s popularity continued to grow strongly, but its culture remains highly polarizing and often generates repetitive online debate. GraphQL is still useful, but mainly for larger/more complex API ecosystems; smaller sites may benefit more from lighter typed alternatives like tRPC. Deno matured significantly with npm support, better editor integration, and real-world production adoption, making it much more compelling than in earlier years. TypeScript has effectively won the typed-JavaScript ecosystem, with near-universal library support and deep VS Code integration. Next-generation devtools are finally evolving after years of stagnation; terminals, code search, and AI-assisted workflows are now more capable and common. The server-side rendering and server-first shift has made working across client and server contexts easier and more practical. Some long-promised technologies, like Temporal and Rome, remain promising but have not yet reached broad everyday adoption. Payment and subscription platforms are often judged on operational economics; the hosts emphasize that build vs buy decisions depend heavily on fees, PayPal support, and control over revenue.

Data Points: Tailwind npm installs (January): 1.75 million - Compared to current installs mentioned during the episode to illustrate growth Tailwind npm installs (current): 4.5 million - Used to show Tailwind more than doubled in popularity Gumroad fee change: 2.7% + credit card fees to 10% - Discussed as a significant pricing increase affecting creators Chipolo One Spot price: $12 each - Presented as a cheaper AirTag alternative AirTag price: $29 - Used for comparison against Chipolo trackers Lemon Squeezy fee: 5% + 50 cents - Mentioned while discussing alternative payment/subscription tooling Temporal status: Stage 3 - The hosts note Temporal is close to standardization but not fully shipped Firefox container queries: Firefox 109 - Mentioned as the version where container queries became available Sentry promo: 2 months free with code 'tasty treat' - Sponsor mention during the episode Sanity promo: Double the free usage tier - Sponsor mention during the episode Auth0 promo: a0.to/syntax - Sponsor mention and call to action Level Up promo: 30% off annual subscription with code 'syntax' - Mentioned during the SvelteKit course plug

Pivotal Quotes: "It has everything I want. And it's all just in this, like, it's very simple, beautiful API." — Scott: Describing SvelteKit after reading the docs and comparing it to other frameworks "TypeScript won, it definitively won." — Wes: During the discussion of typed JavaScript ecosystems and library support "We try to implement all low-level APIs in Rust." — Wes (quoting Jared from Bun): Used to explain why Rust is increasingly important in runtime and tooling internals

Implications: For listeners, the episode suggests that 2023 will likely continue the shift toward typed, server-first, and Rust-influenced tooling, with SvelteKit, Deno, and CSS primitives gaining more ground. It also reinforces that simpler, framework-agnostic libraries and careful build-vs-buy decisions will matter more than hype.